To Rescue a Beached Whale: The Translation of Matthes’ Bugis Dictionary

Campbell Macknight1

Anyone with a serious interest in the Bugis language soon comes to confront the

Boegineesch-Hollandsch Woordenboek of Benjamin Frederick Matthes. This loyal

representative of the Netherlands Bible Society arrived in Makassar in 1848 as a well-

educated, newly-married 30-year-old, keen to come to grips with the task of translating

the Bible into the languages of South Sulawesi. Along the way, as it were, he also wrote a

good deal about the people and society of the area, but his chief interest was always

linguistic. He collected manuscripts of local literature in the search for an appropriate

style for his translations, and published editions of some of the works to be found in the

manuscripts; he produced grammars of both Makasar and Bugis; and of greatest on-going

scholarly importance, he compiled two extraordinary dictionaries. The first edition of the

Makasar dictionary appeared in 1859 and a much expanded second edition in 1883. The

Bugis dictionary appeared in 1874, with a supplement in 1889. In addition to their

function as steps on the way in the process of Bible translation, these dictionaries are

masterworks of nineteenth-century philology and of the European attempt to comprehend,

catalogue and control knowledge, especially where such knowledge could serve a variety

of colonial purposes.2

2

Title page of Matthes’ Bugis dictionary

By the standards of today, Matthes’ opportunities seem remarkably constrained and

things moved slowly. When he arrived in South Sulawesi at the end of 1848, Makassar

was still a small town, but it took him some six months to get out on his first visit to the

south coast. It was not until the middle of 1852 that he was able to spend two months in

the districts to the north of Makassar, yet these were all areas nominally under Dutch

control. At the end of 1854, six years after his arrival, he moved to Maros about 30 km to

the north, where in the middle of the next year his wife died, leaving him to bring up two

young children. Though he did manage some travel into the Bugis areas in the centre of

3

the peninsula over the next year or so, he was frustrated in his attempt to visit the

important state of Bone before he left Makassar on leave in early 1858. In the three years

he was away in the Netherlands, he saw through the press his basic linguistic work on

Makasar some four years after the materials had been sent back for printing.

During his second period in Sulawesi from 1861 to 1869, he paid more attention

to Bugis matters. He was, at last, able to visit Bone, recently defeated in a major military

campaign and here, as in several later trips, he had the support of his good friend, J. A.

Bakkers, an important government official and, eventually, Governor. It was at this time

too, that he was actively compiling material for the Bugis dictionary and finally began

producing Bible translations in both Makasar and Bugis.

Back in The Hague, in two furnished rooms to begin with and being cared for by

his daughter, Matthes pushed on with scholarship and translation. Among much else, he

saw through the press the Bugis dictionary, the very considerable costs of which were

borne by the Dutch government. This productive period was interrupted, however, by an

invitation from the government to return to Makassar and oversee a school to train local

young men who could assist the purposes of the colonial government. He agreed to go for

four years — together with his daughter — and he seems to have been a success;

translations of school books replaced the Bible for a while. He still took the chance,

however, to collect any manuscripts that came his way.

Matthes was 62 when he eventually returned to The Netherlands in June 1880.

Despite his many academic honours, a very public campaign to obtain a University

position for him came to nothing and he quietly settled down again to the work of Bible

translation. It took another 20 years to produce the printed versions of both Old and New

Testaments in Makasar and in Bugis — and then he wanted to start on the Apocrypha!

That was not to be, however, and he ended his days in Nijmegen where his son had

retired after a career in the colonial army. He died in 1908, aged 90.

4

Matthes was nothing if not dogged — just remember that he translated some of

the less interesting passages of the Old Testament not once, but twice! He also had

method; his published works hang together and he makes many cross-references from one

to another. In a very real sense, he created an intellectual universe with his dictionaries

deriving from his manuscripts, his translations from his understanding of the languages,

and the scope of his interests determined by his contacts and enquiries in the field.

Given the demands of developing his field more or less from scratch — since he

paid slight attention to any work done previously — it is not surprising that Matthes

developed an individual style which runs through all his publications. For example, he

was responsible for working with the printers to develop the exact form of the characters

to be used to represent the Bugis- Makasar script in his publications.3 For the

transcription of this script into Latin characters, he devised a remarkably clumsy and not

entirely adequate system.

The subsequent history of scholarship on the languages of South Sulawesi has

been comprehensively laid out by Noorduyn.4 Matthes’ dictionary of Makasar is now

totally superseded — with gracious acknowledgment of Matthes’ labours — by the work

of Cense and Abdoerrahim which, though still in Dutch, is very much easier to use.5 The

only recent help for the student of Bugis lexicography, however, is a rather slight and

poorly edited dictionary by M. Ide Said, which seems to have been compiled without

significant reference to Matthes.6

Some of the strengths and weaknesses of Matthes’ Bugis dictionary have already

been suggested. Undoubtedly its greatest virtue is the range of words included. Because it

is primarily based on the manuscripts which he had collected and studied, it includes a

great deal of archaic and highly specialised material, even if not everything of this type.7

Matthes believed that the manuscripts set an appropriate tone for his Bible translations.

Given his time in the field, he must also have had a good command of everyday spoken

5

Bugis, and it would be interesting to know more about the contributions of particular

informants to the dictionary. Actual usage is illustrated with a wealth of examples and

cognates or sources noted in many other languages, especially Makasar and Arabic, but

also Malay, Javanese and yet more from the enormous range of his scholarship. To

further extend the encyclopaedic range of his work, Matthes made a collection of the

material culture of South Sulawesi as it was in the mid-nineteenth century; he published

illustrations of these items in his Ethnographische Atlas that went through various

editions to accompany the dictionaries. The plates in the Atlas, in effect, illustrate entries

in the dictionaries and they are admirably clear. The items themselves are still preserved

in the National Museum of Ethnology (Museum Volkenkunde) in Leiden.8 The dictionary

also contains a very extensive register of Dutch words with their possible Bugis

equivalents, though this is really no more than a guide to the dictionary proper. Finally,

there are 11 pages of improvements and corrections, to say nothing of another 150 pages

in the Supplement 15years later.

The greatest frustration for the casual user of the dictionary is that the entries are

arranged, not in alphabetical order, but in the unique order of the Bugis- Makasar

characters. This begins, ka-, ga-, nga-, as with other related scripts, but there are

differences as the sequence proceeds. Then there is the difficulty of Matthes’

cumbersome transliteration that requires various diacritics. More importantly, the glottal

stop, a crucial phoneme in Bugis, is poorly and inconsistently indicated. Despite the

abundance of quotations, the source of these quotations is rarely given, or not given

precisely. Moreover, Matthes writes in a peculiarly ponderous, laboured Dutch, even by

the standards of nineteenth-century scholarship; some of his entries could be charitably

described as less than lucid. Finally and most distressingly, the publisher chose to use a

paper that has degraded badly with the years; pages flake and tear with even the most

careful handling.

6

Here, then, is a true leviathan of scholarship, but in this day and age, cast up on a

rather lonely beach. How many people can use the guidance of this mighty whale in order

to launch with confidence into the ocean of Bugis language and literature? In particular,

the product of Matthes’ labours is effectively denied to those young Indonesians who

might best profit from it.

<h1> The Project: Rescue

It is a rash enterprise indeed to begin a new dictionary — and life is short.9 But a new

dictionary is not necessary here. There is much sound scholarship in Matthes that does

not need to be re-done, even if much needs to be checked and, where appropriate, up-

dated. What is needed is a translation, with opportunity for annotations of many kinds.

Almost anything would help at some level.

What follows is a sketch of how a translation could be managed, making use of

contemporary technology and allowing the work to be spread around amongst anyone

minded to help. If money could be found to support particular parts of the project, that

would make things go faster. What are the chances of the project making significant

progress? I don’t know, but the task is certainly feasible, and one can only start by laying

out a plan. If any inspiration were necessary, then Ian Proudfoot’s Malay concordance

project is a marvellous model of what imagination, daring and dedication can achieve.10

The format of the translation would be a website, from which anyone could

download material and, if they wish, print off a hard copy version of material. That pre-

supposes finding a server which is likely to remain viable in the long term, and there are

some obvious institutions to approach. Given that this would be an evolving website, I

doubt the need for a print-on-demand facility, but it is hard to know how the technology

will change. More important is to establish the principle of free and open access, even if

some restriction would be essential for making changes to the website. To ensure

7

consistency and standards, there would need to be an editor, or at least a very restricted

editorial group, who would check and integrate the work of contributors. There is a strong

case for using a data-base programme for the translated material and professional

linguists have developed programmes that would be suitable, but there is also a case for

keeping everything very simple and using Word, as in the examples below. At its most

basic, the website would give access to a series of PDF files, each developed from a

masterfile and replaced from time to time as new work was done, checked and entered

into the whole. The website would be divided into five parts.

Part A would consist of introductory material explaining the project and setting

out the contents of the other parts. In its initial version, it would be wholly in English, but

perhaps an Indonesian version might be added one day. It would be worth including here

translations of Matthes’ prefatory material to the dictionary, the supplement and the

ethnographic atlas since this provides the context of the work and some useful

information.

Part B would consist of digital images of the pages of the three works: dictionary,

ethnographic atlas and supplement. The files for the dictionary, which runs to a total of

1,188 pages, would need to be divided to allow easy navigation, but there are no technical

problems with this. The National Library of Australia has recently made available a

digital version of R. J. Wilkinson’s Malay-English dictionary of 1903 and this shows how

such a project can be done.11 Given the relative rarity of the items — and even more the

fragility of the pages of the dictionary — there are preservation grounds for this exercise

alone, but for the translation project, it is also important to have easy access to the

original materials.12

Part C would be the translation proper. For ease of control and access, this would

be divided into about 30 sections, each containing a few hundred entries, and the exact

division between sections would be governed by how the totals fell out for particular

8

letters of the alphabet. Each section would indicate the date and stage of its latest

revision. This matrix of division and stages would allow the total work to be managed as

much smaller tasks and, as a new stage was completed for a particular section, this text

would then replace the previous version of the same material on the website.

Stage 1 would consist of no more than a list of head-words or lemmas, taken

directly from the dictionary but re-ordered alphabetically. This list would provide the

skeleton of the divisions on which all else would rest. Although new entries could always

be introduced, additions would be rare and would not upset the division into sections.

Stage 2 would be concerned, essentially, to identify each head-word with a

simple gloss of one or two English words. Where Matthes distinguishes more than one

basic meaning for a single head-word, which he marks by numbers, these separate

meanings would also be distinguished at this stage.

Stage 3 would introduce some, at least, of the cognates provided by Matthes and

the derived forms he lists, in a version of his transcription. Some of the cognates could be

usefully checked at this stage, especially the Makasar cognates against Cense and

Abdoerrahim’s dictionary and the Malay cognates against Said’s Bugis-Indonesian

dictionary.

Stage 4 would involve a careful revision of the English glosses and entry of the

examples of usage given by Matthes in Bugis and English translation. A matter on which

Matthes’ transcription is inadequate is in the identification of various prefixes, and that

would require particular care at this stage. Without going into the vexed question of the

best Latin orthography for Bugis, it would be important to convert all Matthes’ Bugis to a

modern standard. This would also be the best stage to note references to the ethnographic

atlas and a few other sources mentioned by Matthes.

Stage 5 would allow for the entry of material beyond that found in Matthes. For

example, it would be useful to add any further Indonesian glosses, mostly to be found in

9

Said’s dictionary. This extra material would need to be carefully distinguished and

identified. Where there are difficulties, it would be possible, in this format, to have quite

extensive discussion. Several possibilities for such additional material present

themselves: the meaning of some words will have changed over the period of more than a

century since Matthes was working and modern Bugis speakers might wish to comment;

anyone working on Matthes’ manuscripts and published materials will quickly find

usages quoted without attribution and these could be identified;13 and finally there are

words, especially in La Galigo and other obscure texts, where there is scope for the

discussion of meaning beyond the understanding Matthes had. This final stage would

remain open-ended.

Part D of the project would take up one of the most useful aspects of the

dictionary, that is the register of Dutch words. It would, in theory at least, be easy to

develop a table of four columns listing equivalent Dutch, Bugis, English and Indonesian

words from the glosses at various stages of part C. There would be, of course, many

problems of detail, but the usefulness, even of an imperfect attempt, justifies the effort.

With the ability to sort these columns alphabetically, one could present three versions of

the same material arranged to allow access from Dutch, English or Indonesian.

Finally, part E would list references, abbreviations and any other peripheral

material.

<h1> The Translation

Perhaps this is all just a castle in the air, but it is possible to venture a sample of the

dream. In particular, the attempt allows experiment with conventions of font and

presentation. I tend to favour the use of Garamond as a font, and somewhat to my

surprise, find Matthes’ basic style of roman for words in Bugis, and italic for words in all

other languages, the clearest. He is not, however, entirely consistent.

10

What follows is an attempt to show what the process of translation would look

like. These six entries are drawn from pp. 32–3 of the dictionary and represent the variety

of entries. They show how many of the issues mentioned above could be dealt with and

the notes draw attention to matter of special interest. In order to minimise repetition, it is

only necessary to show stages 2 and 5 of the process outlined above.

Pages 32 and 33 of Matthes’ Bugis dictionary

11

<h1>Stage 2 16.1.07

<h2>KERRANG 1. Stand on end

2. Open

<h2>KERRÉ’ Phlegm

<h2>KERRI’ Scratch

<h2>KERRING Violent

<h2>KERRO 1. Court song

2. Profane

<h2>KERRU14 Clucking

<h2>Stage 5 19.1.07

<h2>KERRANG 1. To stand on end, used of the hair of the head or

body hair in general

From there, kerrampulu a sort of charm or spell that strikes anyone with fear

and awe of its owner and, as it were, makes the hair of their whole

body stand up from fear and alarm

[Jonker notes a use of the word in Matthes’ Boeginesche Chrestomathie vol. 2, p.

273, line 9 up (actually it’s 11 up!). Another notable use of this term

is in the nickname of the third ruler of Boné, Kerrampélua’, whose

hair is said to have been standing on end when he was born. See

12

under wélua’. The supplement adds a Mak. cognate korang; this

appears in M’s Makasar dictionary, but not in C.15]

2. To open, open up, make loose

Mak. karang

For example, an umbrella or sunshade, sails, etc.

<h2>KERRÉ’ Phlegm

Mak. karra’

Makkerré’ to bring up phlegm, to clear one’s throat, to vomit

<h2>KERRI’ Scratch

Mak. kari’ M schrab, kras16 S kukur

Makkerri’ to scratch, to scrape off; Puru-makkerri’, see puru 1; Kerri’i berre’é to scrape a

stick across the top of a measure of rice (to check that it is not over-full)

Pakkerri’ the stick for scraping; Pakkerri’-kaluku, the Mak. kikkirang, an oblong piece of

timber with a piece of iron projecting from one end which has an point

like a riding spur and is used to scrape out coconuts. (Eth. Atlas pl.

11/25) [The illustration is very clear. The object itself is series 37/item

88 in the museum.]17

Pakkerri’-pe’jé a hook used to scrape the salt together (in the saltpans).

Pakkerri’-kanuku one of the roe-roe [ornaments] used to file the finger nails

<h2>KERRING Violent, strong (e.g. used of an east wind).

Mak. karring M hevig, sterk

[C gives the sense of dry, with the Ind. cognate kering.18]

S taggerring-kerring berdebar-debar19

13

<h2>KERRO 1. A kind of court song for births, circumcisions, etc.

Mak. karro

N.B. the book in which the kerro is written is also called kerro; from this, riwata’ni

kerroé the kerro is brought up in the house, see wata’

Makkerro to chant the kerro

Toriakerroang he or she for whom the kerro is chanted

Pakkerro those who chant the kerro

2. Profane, sacrilegious

Cf. Arab. makrūh Mak. karro

For example: makkerroni puasaé the fast is violated, Harangngi makkerrotoi it is

prohibited and sacriligious

<h2>KERRU Clucking, the sound one makes to call chickens

Mak kurru Ind. Jav. kur Bat. hurho

[The pronunciation of this word is variable. Murni Mahmud prefers kuru’.20]

From this it is used metaphorically: makkerru-sumange’na21 to summon someone’s

sumange’ or personal life spirit (see under sumange’) as if it were a bird,

especially a hen; this may be to prevent it flying away or, to make it

return when it might already have flown away. Thus one says, for

example, kerru-sumange’mu summon your sumange’, my child! that is, so

that it should not depart. This expression is used for a child that shows a

special sweetness, since one presumes that the evil spirits might regard so

much sweetness with envious eyes and on account of this bring harm to

such a child just where they can. With this expression, therefore, one

indicates indirectly that the child has a lovable appearance, and it is

14

preferable in the eyes of the native to express it like this, rather than to

say straight out macapadé-padémani, or such like (see below), since this

would cause the spirits to be slightly more envious.22

Mappakerru-sumange’ to scatter cucubanna and wenno (coloured rice and roasted

material; see under these terms) over someone in order to summon his

sumange’, that is, in order that it should not depart. N.B. this happens in

respect of bridegrooms and all high-placed people when one welcomes

them, anxious as one is that the evil spirits out of envy at the joy of their

high rank might cause their sumange’ to depart.

Pakkerru-sumangekkeng something or another (a buffalo, for example) used for someone to

summon his sumange’. For example: Narékkua arungngé mappégau’-gau’,

namasélemmanenni tau-maégaé, upakkerru-sumangekkengngi if the ruler

gives a feast and thereby does something to frighten all the people away, I use (that is,

sacrifice) (one black buffalo) in order to summon back the sumange’ which has

already been driven off by fright (or, was on the point of flying off). [A

simpler translation would be: If the ruler gives a feast and all the people are

shocked, I would restore confidence.]23

Pappakerru-sumange’ literally a summoning of the sumange’; from this, a means of

preserving the sumange’; from this, a present of a kris or a pike or other similar

object.

15

Notes

1 I would like to thank Ian Caldwell, Jason Lee Kwok Loong and the editors of this

volume for their helpful comments on a draft of this paper.

2 The two major sources for Matthes’ life and work are H. van den Brink, Dr Benjamin

Frederik Matthes: zijn leven en arbeid in deinst van het Nederlandsch Bijbelgenootschap

(Amsterdam: Nederlandsch Bijbelgenootschap, 1943) and J. L. Swellengrebel, In

Leijdeckers Voetspoor I 1820–1900 (’s-Gravenhage: Nijhoff, 1974), pp. 147–72.

3 See J. Noorduyn, ―Variation in the Bugis/Makasarese script‖, Bijdragen tot de Taal-,

Land- en Volkenkunde 149 (1993): 533–70, especially pp. 543–4. The script is best

described as an aksary from the Sanskrit word for character, aksara; the script is based on

the Indic system found widely in Southeast Asia, though its exact derivation and history

are obscure. In addition to the ―standard‖ script used by Matthes, there are many variants,

as well as the so-called ―old Makasar‖ aksary with characters of a quite different form.

4 J. Noorduyn, A Critical Survey of Studies on the Languages of Sulawesi (Leiden:

KITLV Press, 1991). An important recent addition to this survey is Anthony Jukes,

―Makassarese (basa Mangkasara’): a description of an Austronesian language of South

Sulawesi‖, unpublished PhD thesis, University of Melbourne, 2006.

5 A. A. Cense in samenwerking met Abdoerrahim, Makassaars-Nederlands Woordenboek

(’s-Gravenhage: Nijhoff, 1979).

6 M. Ide Said DM., Kamus Bahasa Bugis-Indonesia (Jakarta: Pusat Pembinaan dan

Pengembangang Bahasa, Departemen Pendidikan dan Kebudayaan, 1977).

7 See Ü. Sirk, ―On Old Buginese and Basa Bissu‖, Archipel 10 (1975): 225–37.

Sirk, in his grammar, The Buginese Language of Traditional Literature (of which

the most accessible version is the English translation published in Moscow in 1996),

16

notes that ―it is by far not exhaustive as regards the language of traditional

literature‖ (p. 10).

8 Handling the actual items produces some ironies. For example, a well-made earthenware

jar, shown in the Ethnographische Atlas plate 10, item 9 and rather misleadingly termed a

gandi (kendi), is registered in the museum as series 37, item 35. It is even illustrated with

a good photograph in H. H. Juynboll, Catalogus van ’sRijks Ethnographisch Museum vol.

16 (Leiden: Brill, 1922), p. 20 and plate 5/1 — or on the museum’s website

(<www.rmv.nl>). Surprisingly, it turns out to have a stamp on the base indicating its

European manufacture.

9 James Murray estimated that he could complete the Oxford English Dictionary in ten

years; it took 49. See L. Mugglestone, Lost for Words: the hidden history of the Oxford

English Dictionary (New Haven and London: Yale University Press, 2005), pp. 1–2.

10 See <http://online.anu.edu.au/asianstudies/ahcen/proudfoot/MCP/>

11 See <http://nla.gov.au/nla.gen-an3363573>

12 The Inter Documentation Company has published a microcard edition of the dictionary

and some work has been done to digitise and translate the Ethnographische Atlas.

13 Considerable help with this process of identifying examples is available from an

annotated copy of the dictionary once owned by Prof. J. C. G. Jonker and now held in the

Leiden University Library. There is also a microfilm of this copy in the Australian

National University Library.

14 These have been re-ordered from the dictionary order: kerrang, kerri’, kerring, kerru’,

kerré’, kerro.

15 Note the possibility of adding additional relevant information, especially the Jonker

material.

17

16 It may be occasionally helpful to give Matthes’ Dutch gloss, but this would always be

available in the original materials

17 Note the helpfulness of the Ethnographische Atlas. I have not handled the object itself,

but it is duly listed in Juynboll’s Catalogue van ’sRijks Ethnographisch Museum, vol. 16,

p. 15.

18 C indicates the Makasar dictionary of Cense and Abdoerrahim and would be used

where this supplies significant further information.

19 S indicates the Bugis dictionary of M. Ide Said and would be used only where the

relevant information differs from that in Matthes.

20 I am most grateful to Murni Mahmud for her expert criticism and enthusiastic

assistance, especially with the stage 4 revision of the derived forms and these stage 5

annotations.

21 The use of the hyphen in constructions of this type is debatable. On the one hand, it is

useful to distinguish the two elements; on the other, it keeps them together for further

derived forms as shown below.

22 This rambling explanation is, unfortunately, typical of Matthes.

23 Matthes’ translation shows his habit of trying to explain — often somewhat tediously

— rather than just translating the sense of the words. Swellengrebel, In Leijdeckers

Voetspoor I, pp. 162–4 gives examples of this tendency in his Bible translations. I suspect

that this example is drawn from a manuscript.